Miro Salem, Global Head of Artificial Intelligence & Autonomous Networks, Rakuten Mobile, is speaking on a Panel at FutureNet Asia on 27th October. We recently caught up with Miro to get his thoughts on the panel topic: Intelligent operations for new 5G and Edge ecosystems: A game changer?
How do you start building intelligent operations for a 5G and edge world?
A solid foundation is key. Intelligent operations require the right tools and culture, from efficiently extracting and understanding network data to taking intelligent action. Without a stable ecosystem of tools (e.g., data platform, AI platform, usecase backlogs, etc) and full synchronization across the organization (e.g., RAN, core, cloud, security, devops, etc), intelligent operations will be siloed, duplicated, and inefficient. They require alignment and buy-in from all stakeholders. Only with this clarity can true intelligent operations across various layers and elements of the production network flourish.
Should CSPs adopt a multi-cloud model for delivering immersive, multi-access edge computing services and use cases?
Yes. 5G brings new opportunities and challenges that we must overcome. One of those challenges is the much lower latency requirement all the way to the physical layer. 5G has a much shorter Transmission Time Interval (TTI) than its predecessor, 4G. This means processing and round-trip delay must be carefully considered, especially for mmWave use cases. Pixel streaming gaming is an example application use case that demonstrates the need for edge computing services. More and more applications will need to use edge for low latency (e.g., AR, VR, etc). The need for the multi-cloud model becomes stronger when considering inter-operator edge clouds.
How can CSPs embed the right capabilities and technologies for scalability and automation in their 5G propositions when the business case for growth remains unclear?
The digital transformation of telco has given operators, what I believe to be, the ultimate enabler: virtualization. Embracing the fact that new technologies, services, and applications can be deployed within minutes, even seconds, including in the RAN, we can tailor best-fit solutions and adapt as we go. A fully virtualized, containerized, modular end-to-end network allows operators to respond based on customer needs, not vendor requirements. Like any entity in search of a business model, agility and speed to adapt to customer needs are essential. Business models are discovered, not dictated.
How important is automation’s role in intelligent operations?
The role is fundamental. Automation is the first practical step towards autonomous networks, which is our final goal. Intelligence is a key enabling suite of technologies and mindsets to take the next step towards autonomy. That said, it is important to remember that sometimes automation is sufficient for the task at hand. As with all things in life, it depends on the problem we are trying to solve.
To hear more insight from Miro, join the event and register here